Test participants see a red “Discover” icon marked as new at the top of the screen. The first option inside is organized around moods, so a viewer can pick a theme such as “Scare Me” and get a page of relevant movies and shows. During a media preview, Netflix also showed themes like “Solo Movie Night” and “Swoon-worthy love stories.” Each theme opens a page of titles with art, trailers or descriptions, mixing Netflix originals with licensed content, and subcategory buttons lead into narrower picks such as “offbeat stories” or “workplace comedies.”
Telling Netflix What to Stop Showing You
A “Tune Your Recommendations” button gives viewers two choices on any title:
- Keep recommending similar titles
- Stop recommending it
Selecting “Do not recommend” prompts Netflix to ask why, and those answers further train the AI behind the tab. Netflix says the system is designed to adjust to each individual viewer over time. Participants can choose to skip the feature entirely, and the feedback they give will shape how the tab works before it opens to more users.
The test follows earlier talk of AI recommendations by Netflix execs on an earnings call. The company has been experimenting with AI for discovery for some time. In May 2025, Netflix began beta tests for AI-powered search in its mobile app, which accepted conversational requests like “I want something scary, but also funny.” It is part of a wider push to expand AI across consumer products, and platforms are increasingly leaning on personalization to boost engagement.
Hashlytics Take
The interesting shift is where this lives. The May 2025 test put conversational search on a phone, where typing a request is natural. A TV remote is a poor place for typing, so AI Discover swaps free text for tappable moods and themes. That makes the “Do not recommend” prompt the quietest and most valuable part of the feature: a viewer who explains why they rejected a title hands Netflix a cleaner signal than hours of passive watching ever could. The open question is whether the mood pages feel like curation or like the same catalog with new labels.
Follow Hashlytics on Bluesky, Facebook, LinkedIn , Telegram and X to Get Instant Updates
